strisunshinewentingwang/pytorch

Pull Requests Commits

deduplicate code

izaitsevfb committed 1 year ago

f723ef1d

fix the case when rev-list returns multiple commits

izaitsevfb committed 1 year ago

ef9b433a

izaitsevfb committed 1 year ago

bb4ac7a3

[stronghold][bc-linter] correctly determine the PR base commit

izaitsevfb committed 1 year ago

996f6200

[reland] remove typed StorageImpl::data() and StorageImpl::unsafe_data() (#98411)

mikey dagitses committed 1 year ago

49b80c3e

[dynamo 3.11] fix 3.11.2 issues (#98364)

williamwen42 committed 1 year ago

e6631438

Reduce includes of CUDACachingAllocator.h (#97072)

zdevito committed 1 year ago

1bcb8808

Cleanup Copy.cu logic (#97071)

zdevito committed 1 year ago

e085acc9

[inductor] do not generate loops when the condition doesn't hold (#98185)

nkaretnikov committed 1 year ago

938c5da6

Add max-autotune compilers to benchmarks (#98464)

williamwen42 committed 1 year ago

bb331739

Revert "Move functional collectives implementation to python. (#98315)"

pytorchmergebot committed 1 year ago

67d1a770

Support `getattr` for ConstantVariable when compiling with Dynamo (#98153)

chengzhou9 committed 1 year ago

ce797795

Improve dynamo support for autograd.Function (#98158)

jansel committed 1 year ago

4716fa24

Remove unnecessary schema_map from spmd API (#98444)

mrshenli committed 1 year ago

0c5389b4

[inductor] Enable CudaWrapperCodeGen for non-AOT mode (#98264)

desertfire committed 1 year ago

77f32eb6

[inductor] Combine CppWrapperCodeGen and CppAotWrapperCodeGen (#98088)

desertfire committed 1 year ago

348dcf51

[pt2] add meta function for `take` (#98451)

nkaretnikov committed 1 year ago

7b259763

[Easy] remove unnecessary get_rank() in tests (#98445)

mrshenli committed 1 year ago

01991409

Add new aten::device variant to TorchScript (#97023)

Thiago Crepaldi committed 1 year ago

bbf180af

Improved configuration naming for repetitive workflows (#98496)

DanilBaibak committed 1 year ago

d1e7434b

[Sparse] Raise exception when expand is called on sparse tensor (#98365)

malfet committed 1 year ago

fa4cab89

Move functional collectives implementation to python. (#98315)

Rodrigo Kumpera committed 1 year ago

8b0374f8

Add mark_static (#98427)

ezyang committed 1 year ago

f98c1809

Turn off divisible_by_16 for dynamic shapes; support ablation (#98471)

ezyang committed 1 year ago

bdb79a8f

[quant][pt2e] Support quantizer API in prepare_pt2e_quantizer (#97994)

jerryzh168 committed 1 year ago

3142ce20

[Inductor] Fix convolution lowering if stride or padding or dilation is 1 element list (#98448)

yanboliang committed 1 year ago

ccc27bc3

Print collective (#97544)

rohan-varma committed 1 year ago

b8cf0101

[PG Wrapper] Add sequence number (#97462)

rohan-varma committed 1 year ago

dab1a7e6

[FSDP] records for composable (#98428)

rohan-varma committed 1 year ago

428c531d

[ROCm] Enable FSDP BF16 comm hooks unit tests (#97517)

lcskrishna committed 1 year ago

eadd84d0

Older